Have tclkits died
xvart
for tclkits and the build packages will not compile on fedora..
suchenwi
> Just wondering if tclkits have died off, tcl.tk only has old entries
> for tclkits and the build packages will not compile on fedora..
The problem is more that tcl.tk isn't maintained so often. The latest
news for any Tcl-related topic is usually pn the Wiki:
http://mini.net/tcl/tclkit - and in this case, the project homepage
http://www.equi4.com/tclkit.html
Larry W. Virden
suchenwi wrote:
> xvart schrieb:
>
> > the build packages will not compile on fedora..
>
> http://www.equi4.com/tclkit.html
Has the original poster submitted bug reports to fedora as well as to
the tclkit web sites so that appropriate steps can be taken to fix
things?
Andreas Leitgeb
The original rant was about lack of new tclkits.
Looking at the download-matrix, I see 8.5a4 tclkits,
that are now more than one *year* old.
Considering, that the tclkit-8.5a4 doesn't even
know about the arbitrary-length integers yet,
the OP's worried question is just too understandable.
Eckhard Lehmann
Andreas Leitgeb wrote:
> > suchenwi wrote:
> >> in this case, the project homepage
> >> http://www.equi4.com/tclkit.html
>
> The original rant was about lack of new tclkits.
> Looking at the download-matrix, I see 8.5a4 tclkits,
> that are now more than one *year* old.
>
At the download matrix (http://www.equi4.com/pub/tk/downloads.html) I
see Tclkit-8.5a4 from 19/04/2006, which is just around the time when
tcl-8.5a4 was released.
Eckhard
Andreas Leitgeb
> Andreas Leitgeb wrote:
>> >> http://www.equi4.com/tclkit.html
>> The original rant was about lack of new tclkits.
>> Looking at the download-matrix, I see 8.5a4 tclkits,
>> that are now more than one *year* old.
> At the download matrix (http://www.equi4.com/pub/tk/downloads.html) I
> see Tclkit-8.5a4 from 19/04/2006, which is just around the time when
> tcl-8.5a4 was released.
Oops, my mouse hovered above the "xft"-version of tclkit.
Indeed, the other version appears newer at first glance, but
then I downloaded, ungzipped and tried it:
$ ./tclkit-8.5a4
% expr 2**70
0
%
So, no matter what the exact age, it's *far* behind
current head, so the worries still apply.
PS: Of course, I know, there exists no *right* for new versions,
so what I do here is not "ranting" but rather "mourning" ...
Don Porter
> $ ./tclkit-8.5a4
> % expr 2**70
> 0
> %
>
> So, no matter what the exact age, it's *far* behind
> current head, so the worries still apply.
That bug fix is two months old.
So, to clarify, you expect contributors of built tclkits to contribute
an updated one more frequently than every two months? What frequency do
you require?
--
| Don Porter Mathematical and Computational Sciences Division |
| donald...@nist.gov Information Technology Laboratory |
| http://math.nist.gov/~DPorter/ NIST |
|______________________________________________________________________|
Andreas Leitgeb
It seems, now it's time for me to apologize...
I didn't know/think about this particular bug, and
expr 2**70 was just my quick-test to try to generate
some big numbers. Had I tried 3**70 instead, the last
few postings would have been avoided *shame on me*...
> What frequency do you require?
The current frequency seems ok, then.
My *wish* would be one release after each "biting" bug has
been fixed (for some threshold of "biting"), but well,
its 0-cost software, and all that ... :-)
Larry W. Virden
Andreas Leitgeb wrote:
> My *wish* would be one release after each "biting" bug has
> been fixed (for some threshold of "biting"), but well,
> its 0-cost software, and all that ... :-)
I suspect that the release rate is "one per alpha release, plus
possible updates if the build of the tclkit itself is in some way
defective". Since 8.5a5 is still in process, no new updates have
occurred.
Larry W. Virden
Andreas Leitgeb wrote:
> My *wish* would be one release after each "biting" bug has
> been fixed (for some threshold of "biting"), but well,
> its 0-cost software, and all that ... :-)
Hmm - I wonder - is it possible to set up sf.net resources to generate
the tclkits on a "as updates are checked in" basis, then to run a test
suite against it? Sort of a "continuous build" environment? anyone
want to look into that - in theory, the sf.net compile farm should be
able to create a variety of executables. And if the software were built
daily, with test suites run and reports generated, then when the
release manager was ready for a release, it would be a matter of
selecting the appropriate binary.
Eckhard Lehmann
Larry W. Virden wrote:
> Hmm - I wonder - is it possible to set up sf.net resources to generate
> the tclkits on a "as updates are checked in" basis, then to run a test
> suite against it? Sort of a "continuous build" environment? anyone
> want to look into that - in theory, the sf.net compile farm should be
> able to create a variety of executables. And if the software were built
> daily, with test suites run and reports generated, then when the
> release manager was ready for a release, it would be a matter of
> selecting the appropriate binary.
That's what I asked myself in this moment. Automated nightly builts of
Tcl, Tk and Tclkit would be very nice... not for official production
releases but for testing the "bleeding edge" :). AFAIK, Mozilla is
doing this for its product suite.
It shouldn't be complicated - a script that does cvs update of a local
working copy, makes the built/distro and copies the result to an
appropriate place. Plus an entry in crontab (or whatever job scheduling
mechanism) to run it frequently...
Eckhard
Larry W. Virden
Eckhard Lehmann wrote:
> Automated nightly builts of
> Tcl, Tk and Tclkit would be very nice... not for official production
> releases but for testing the "bleeding edge" :).
if you are familar with setting this up, perhaps you could contact the
TCT and volunteer to set up and babysit stuff...
Uwe Klein
uwe